Workday Podcast: How Comcast Turns Data Into Insights With Workday Prism Analytics

Jeremiah Barba:很少科技和媒体公司的领导人would tell you that they are suffering from a shortage of data, especially in HR. But what to do with that data is an entirely different story, and getting real insights from it is another challenge entirely. I'm your host, Jeremiah Barba, and today, I am joined by Jay Chang. Jay is the executive director of HR analytics at Comcast, and he's here today to talk about how they use Workday Prism Analytics to turn data into insights. Jay, thanks for joining me today.

Jay Chang: Thank you so much for having me.

Barba: Tell me a little bit about yourself and what you do at Comcast.

Chang: Sure. I've been with Comcast since 2009. I started out as a compensation director. Generally speaking, in HR, the, the quantitative folks end up being comp. And so I did that for a few years. I got to spend some time in California as a result of that. And then I came back to Philly, and I started doing more generalized HR analytics. So stuff around turnover, stuff around hiring, cost of hire, those kinds of things. And then I got involved with the Workday implementation from a reporting perspective. And so I spent, I want to say two years, three years - I actually lost track basically helping to build out all of the reporting and, dashboarding and such that was going to go as part of our implementation tenant. So we went live in 2021, and we've been there ever since.

Barba:让我们开始with a why question, sort of set the stage. Why did you choose Workday Prism Analytics at Comcast? And how did that go hand in hand with Comcast's overall goals?

Chang: We were starting with a greenfield, right? So we had SAP previously, and we were moving from an on-prem solution to the cloud. And we looked at a couple of other options. And the Workday solution was chosen after a really long evaluation process. And we decided at the time that we needed something above and beyond what was in standard Workday reporting. The Workday product as a whole has a lot of reports. But when you're a company of our size, we're used to getting stuff in our own unique ways. And even the same report type in one group is completely different in terms of fields and layouts and those kinds of things, for another group. And so to kind of support that, plus the fact that we knew that there was a whole bunch of data that we were going to have to take in from other sources or send out to other sources, Prism was a natural fit. And then, we were also using a data warehousing solution from SAP, and that had to be cared for from a replacement perspective. And so Prism was kind of sold to us as the solution for trying to manage the data warehousing needs that we were basically coming from a legacy data perspective.

Barba: No, that makes sense. That's, that's helpful background. So I understand that you, you implemented Workday, I guess, quote-unquote, "out of the box." Could you share a little bit more about how you approached that implementation and, and what you learned from it?

Chang: We really like to say that we try to do everything out of the box. And it's hard to tell because Workday is, you buy it, and then you customize it in the sense of basically saying, "Okay. I want this process to work this way. I want that process to work that way." So it's technically out of the box, but you've customized it to your own process. We like to say that it's fairly standard. And I think for the most part it is. But at the same time, we do have a lot of custom reports that are running around in the system. And I think everybody who does Workday for any period of time will have that, right?

Barba: Sure. It sounds like you took advantage of the flexibility to make it what you needed, right?

Chang: I think that's what is very true today is that when we started with Workday, we had the team that did the implementation. They lived through that whole process. But then when we actually went live, that's when you really try to start figuring out, "Okay. Now, this is how-- what this actually means." And as we got better at it, as we got further along the process, we started realizing, okay maybe this would be better and so we need to change it, right? So now we're going through that process of refinement. And I think that's pretty natural.

Barba: So in your session here this week, you talked about how you approach different use cases for Workday Prism Analytics and their varying levels of complexity. Could you talk a little bit more about that?

Chang: Sure. So regarding Prism, is, at the end of the day, a, a tool to load data, do stuff to it, and then join it to Workday data and publish it into Workday for reporting purposes. So the most simplistic use case is, "Hey, I have this external file. I get that file on a regular basis. I need that joined to employee data or compensation data or whatever." And so I can import that data through Prism into Workday. I can join it with whatever else I need to join it with. I can secure it however I need to secure it. And then I publish it to Workday to create a report against it, right? So that's the most simplistic use case. So, there are cases that we have where we are taking compensation data, and we are basically transforming it so that we have a version of everybody's pay for every pay zone that we have in the company. And what that allows us to do is compare the same job across the entire population of people that have that job regardless of where they are in the country, right? And so that allows you to say, "Okay. Well, this director of government affairs is being paid X, which is 35% in range." But that 35% in range could vary slightly whether you're in San Francisco or whether you're in, you know, Des Moine, Iowa or Washington, D.C. And it has historically been very hard for us to compare that, right? Because generally speaking, companies pay a little bit different depending on where you are in the country.

Chang: And what this approach does is it says, "Okay. Well, we'll pretend that everybody works in the same place. And we will scale their salary up or down relative to where the job is so that you can say, 'Okay, generally speaking, a director of government affairs is at 37% in range.'" And then that helps inform you as to what is the right salary to offer a person, right? And then you can also see it within the context of how long have they been in the job, how long have they been with the company. These things kind of effect the overall offer that you want to make to the person to be equitable to both the person and the peers that are there in the job. So that's one example where we're taking data that isn't necessarily outside of Workday, but we're pulling it into Prism, we're doing some stuff to torture it, and then we're sending it back. It’s also great for complex business logic. I have another use case where we are basically taking leave data, and we're trying to figure out, "Okay. Jay's been on six different leave types given how leave works, some of them overlap, some of them run concurrently, some end before others start? And so we're trying to figure out, "Okay. At the end of the day, how long has Jay been out of work?" Right? And from a management perspective-- of course, we as HR care about the type of leave you're on, whether or not you're allowed, whether or not you're paid or not, paid, you know, how long you can be on that leave, and whether it's protected or not. Okay, those are all HR things. But at the end of the day, the business, they just want to know, "Can Jay work?" Right? from their perspective, it doesn't matter to them how many different leaves Jay is taking on or being awarded because they expect HR to do that. That's, that's what HR does. But from their perspective, they want to know, "Okay, Jay went out yesterday. When can I expect Jay back?"

然后杰继续他的叶子,他们想要的东西to get a sense, "Okay. How long has Jay been out? Not necessarily on leave type A or B, but how long has he been out, and when do we expect that that might--- that might change?" So those are two examples of relatively medium complexity logic. And then I'm going to talk about a couple of, use cases where we just go crazy with data, right? Like, so, Workday has an object called trended teammate-- or trended worker, which is their way of doing historical teammate reporting. And we looked at that, and so now what we're doing is we're building our own version of that. And that is basically allowing us to track things like hiring and terminations and movement. But it's proving to be incredibly helpful for other things. So now, we took that data set and we are combining it with our recruiting data so that we can see, "Okay. Here's a fill. This is the person that is tied to that fill and all of the data associated with that person." Likewise, we are looking at it from a succession perspective, and we can say, "Okay. Well, here's a pool of succession candidates. Here are the jobs that they have taken over the past X number of months." And we can tell were they or were they not the successor for those jobs. And by pulling that data into Prism and then by joining it with the data that's sitting in our version of trended worker, we're able to do a lot of stuff that, that ties back basically to the root data behind HR, which is person, hires, moves, terms.

Barba: It's so interesting to see how, like I said before, you have created this flexibility, taken advantage of what Workday was in the beginning, and then you're creating real-life answers to these complex questions through data. Once you start talking about it, it kind of blows my mind, you know, because I'm not an HR analytics expert. So it's incredible to hear how the things that you're working with, the reports, they end up giving you such valuable real-life information.

Chang: Yeah, and for our HR people, there are times when they do need the actual Workday reports and the actual data that those Workday reports return. But there's also a lot of times where it's easier for them to go to one place. And so we have this big dashboard. It has all of these different charts, and they can drill into the detail behind each of the activities. And that way, they're going to one place, and they can see most of what they need to see. And then if they have questions, they can always drill deeper into the individual employee or into a specific report. But a lot of times, 80, 90 percent of the time, the detail that's our movement dashboard is, is really enough for them, right? And that way, it simplifies their lives, you know? And that's really, at the end of the day, what you want is HR needs to become more strategic. We hear that all the time. But you can't make them more strategic if you keep giving them more, like, transactional tasks to do. And so we keep trying to find ways to remove some of the transactional work and some of the operational reporting type of work so that they can spend more time being truly consultative, strategic, and the type of HR person that we really all want to have in our organizations.

Barba: So let's move into this question about the implementation. So what are some of the key lessons that you've learned through this implementation, and you've talked a little bit about this already - how Workday Prism Analytics works together with the rest of Workday? So a two-part question. What lessons did you learn in the implementation? And then what lessons did you learn after the implementation? And how were those different?

Chang所以他们一起的模糊。我将具体讨论cally about reporting and about Prism. From a reporting perspective, if I could tell my previous self, I would push back a lot more on the use of standard reports. Because we're used to how things were in our old system, and now we're blowing everything up. It's not even like we're taking the same process, and we're moving that into the cloud. We actually revamped all of our processes too. So it's very hard to figure out, "Okay, in this new process, what data do you need to support that process? And what reporting do you need to support that process?" And so the natural inclination is to use the Workday stuff because they've already done this thousands of times. And so why would you really think that you could be better than the thousands of times that they've done it before? But I think, as organizations get bigger, the needs of specific groups within that organization change, and it's very hard for those organizations to conform their existing processes to a new way of standardized data view. And so if you have a large organization, if you have lots of complex stakeholders who are each torturing data in their own way, then you, you really do need to challenge anything where they say use the standard reports.

From a Prism perspective, I would tell myself not to be so scared. Because I didn't know what the heck I was doing, you know? And, and so, we had a couple of Workday consultants doing the Prism builds. And they did a lot of them. And when they left, I was so scared to touch anything. So, what I would tell myself is be a little bit more brave and go into the sandbox and go into your development environments and just play around with stuff and figure out what's going on and how this stuff works. It gives you an incredible amount of flexibility that once you start understanding when do I use Workday, when do I use Prism, how do I combine the two, it becomes a very useful tool. And it's a great thing to have in your back pocket if you know how to use it well.

Barba: So it sounds like we have a common theme, right? Taking the flexibility of Workday, taking what you're given out of the box, air quotes, and then just being honest and creative about how you can then, you know, tweak it and modify it and build what you need.

Chang: Yeah.

Barba: So that sounds like that's a great lesson I think for everyone that's involved in these types of projects.

Chang: I mean, you just said a word that I think perfectly sums up the really good Prism consultants and Prism practitioners is creativity.

But you learn to take a set of tools and apply them in ways that perhaps don't necessarily look straightforward. But the more creative you are with those tools, the more complex and, quite frankly, interesting things you can build off of those tools. So I think the people who really enjoy Prism, who really, do well with the product are people who really enjoy being creative thinkers.

Barba: So a good segue to our last question which is around what advice would you give to tech and media company leaders in a position like yours if they really want to focus on improving their analytics in, say, the next 12 months?

Chang: Sure. So there's HR, and then there's analysts, and then there's data. And generally speaking, HR needs better training, better support in understanding data that their analysts are giving them because there are times when people in HR just scared of data. They’re scared of the numbers. They don't know how to read them. They don't understand how they're created. And so because of that, they, they tend to try and avoid it, or they just rely on, "The numbers say what the numbers are because the analyst took care of it." And at the end of the day, as a business partner, as a person who is advising the business, you have to kind of understand what it is that the business is facing from a challenge perspective and then figure out, okay. What's the human component of that? And how do I get to that human component and show the business what it is that is driving some of the, the things that they're seeing? And so we don't, I think, spend enough time helping HR become familiar with data, become comfortable with data. And I think that's something that we need to do better. And I would always advise that people spend time educating the HR business partner on what it means to understand data.

从整体数据的角度来看,我知道我们谈了lot about HR data and a lot about how great HR data is. HR data is not that big. You're not getting promoted every other month, and your compatriots aren't getting promoted every other month. Or maybe they are and you just don't know it. But at the end of the day, if I get one promotion every three, four years, you know, that's probably doing okay. And so from an overall perspective, the amount of data that we have in HR is not huge. So we shouldn't be afraid to kind of look for simple techniques rather than the latest ML, AI, whatever it is that's being thrown out there. Those things work in certain arenas, but for the most part, the day-to-day people in business care about --- when is my person going to start that I hired? Why did this person leave? How can I promote better communication in my team? How can I get my people promoted and recognized? Those are all standard and relatively simplistic datasets, right? So generally speaking, I think we try and make the HR data a little bit more complex than it needs to be. And I don't think that, that that's really necessary. I think there's a lot of value in interpreting the data we do have before we try and build on things that, that talk to data that we might have.

Barba: That's great. Again, thank you. So many good real-life examples of how you're creating value through Workday, through creativity, and, thinking outside the box, if you will, to use a cliche. And again, thank you so much for joining me.

Chang: Thank you so much for inviting me. It was a pleasure.

Barba: Of course. We've been talking about how Comcast finds insights with Workday Prism Analytics with Jay Chang. Be sure to follow us wherever you listen to your favorite podcasts. And remember, you can find our entire podcast catalog at workday.com/podcasts. I'm your host, Jeremiah Barba. And I hope you have a great Workday.